Carole Bewley LBC, NIDDK, NIH Annual Report, Fiscal Year 2002 The Natural Products Group of LBC, NIDDK, is active in two areas of research, both of which seek to identify and understand the structural basis and mechanisms of action of inhibitors to (1.) mycothiol-dependent biosynthesis and detoxification in mycobacteria, and (2.) HIV-1 Envelope-mediated fusion. 1. Glutathione is the primary low-molecular weight thiol responsible for maintaining a reducing intracellular environment and in protecting eukaryotes and Gram-negative bacteria from oxidative stress. Actinomycetes, on the other hand, which include pathogenic groups such as mycobacteria adn streptomycetes, do not produce glutathione, and use instead the small-molecular weight thiol mycothiol (MSH). Four enzymes carry out MSH biosynthesis, while MSH-S-conjugate amidase (MCA) facilitates MSH-dependent detoxification. Three of the four Mtb biosynthetic enzymes as well as MCA have been identified recently, and MSH-deficient mutants display either reduced growth rates or non-viability. Thus, these enzymes represent potential targets for new classes of Gram-positive-specific antibiotics or antituberculars. As part of our ongoing efforts to identify new antibacterials, we have isolated and characterized 13 natural product inhibitors of the detox enzyme MCA, most of which are lethal to M. smegmatis. The chemical structures of our inhibitors led to screening of a synthetic bromotyrosine natural product-inspired library which revealed key structural features lending to good inhibition of MCA. The structures of the natural and synthetic inhibitors suggested MCA and its homolog GlcNAc-Ins deacetylase are metalloenzymes, and we have demonstrated this using standard biochemical techniques. In the course of this work, we also discovered that the absolute stereochemistry of MSH had never been unambiguously established. Thus, we completed a total synthesis of the fluorescent bimane derivative of MSH, namely MSmB, as well as both isomers of the pseudodisaccharide portion of MCA. These syntheses provided the groundwork for our current activites which include the design and synthesis of substrate-based inhibitors that incorporate key features of the natural products and the substrates. 2. In order for enveloped viruses to infect cells, membrane fusion must occur. In the case of HIV, the events leading to fusion include binding of the HIV envelope glycoprotein gp120 to the primary receptor CD4, and subsequent binding to the chemokine receptors CXCR4 or CCR5, which together facilitate insertion of the fusion peptide of gp41 (i.e. the transmembrane subunit of the HIV envelope protein) into the host cell membrane, ultimately leading to fusion. Current anti-retrovirals target exclusively HIV protease and reverse transcriptase (RT). While these therapies have vastly improved the lifestyle for HIV-positive people in developed countries, these drugs do not eradicate the virus. Thus, there is a continuing need for new classes of antivirals. The second aspect of our research focuses on inhibitors of viral entry, and encompasses the following projects: (i.) identification of small-molecule inhibitors to HIV fusion, (ii.) determination of the structural basis of the potent fusion-blocking activity of cyanovirin-N, and (iii.) the design of proteinaceous inhibitors and/or antigens of HIV fusion. i.) To date potent inhibitors to viral entry by HIV are limited to peptides or proteins comprising sequences from HIV-Env, to chimeric proteins incorporating neutralizing antibodies, or to compounds of modest molecular weight (1-5 kDa) that bear an overall abundance of negatively charged moeities. In the 3rd example, binding to Env is thought to be nonspecific and to occur through interactions with the positively charged coreceptor-binding regions on gp120. Thus, the impetus for attempting to identify small-molecule inhibitors to HIV fusion stems not only from the desire to identify new classes of potential antivirals, but also to reveal novel scaffolds that mimic naturally occurring proteins. Earlier we identified five novel bishomoscalarane sesquiterpenes and several pyridoacridines that inhibit HIV-1 Env-mediated fusion in vitro at low to sub-micromolar IC50 vlues. During the past year, we have isolated two new guanidinium alkaloids that inhibit HIV fusion and appear to block gp120-CD4 binding. During the next year, we plan to elucidate the structures of these complex natural products and map their binding site on gp120 if sufficient material can be obtained. ii.) Cyanovirin-N (CVN) is an 11kDa cyanobacterial protein that potently inhibits (IC50=1-10nM) all strains of HIV and SIV at the level of fusion via high affinity carbohydrate-mediated interactions with gp120. As an extension of our earlier findings that CVN is selective for the D1D3 isomer of oligomannose-8 (Man8 D1D3) and oligomannose-9 (Man9) and binds these carbohdyrates with nanomolar affinities, we solved the solution structure of a 1:2 CVN:Mana(1-2)Mana complex using multi-dimensional heteronuclear NMR to reveal the first structure of a complex of a mannose-specific carbohydrate-binding protein with nanomolar affinity. The presence and location of two opposing non-overlapping carbohydrate binding sites in conjunction with the geometrical restraints imposed by the spacing of the D1 and D3 arms of oligomannose presents a simple model for CVN-gp120 binding in which CVN may cross-link individual monomers of gp120. Owing to the putative divalent nature of CVN-gp120 interactions, we went on to engineer an obligate 3-dimensional domain-swapped dimer of CVN by deleting one non-critical residue from the hinge linker about which domain-swapping occurs. We characterized the mutant using analytical ultracentrifugation, NMR relaxation parameters, size exclusion chromatography, and an HIV-1 Env-mediated fusion assay to demonstrate that this mutant cannot exist as a folded monomer. In regard to drug development and design, this mutant exhibits greatly improved a number of significant advantages over wildtype CVN. These include increased stablity, a 5 to 10-fold increase in potency, creation of a novel quadravalent carbohydrate-binding protein, and argubly of most importance, a molecule that can be purified to greater than 95% homogeneity from a crude E.coli cell lysate in a single chromatographic step. Last, we also demonstrated that the low affinity carbohdyrate binding site of CVN becomes a high affinity carbohydrate binding site for oligomers bearing the structure Man-alpha(1-2)Man-alpha(n). Although such linkages are not found in mammalian carbohdyrates, they are abundant in many microbes, including the LAMs of mycobacteria, suggesting that CVN plays a broad defensive role in its cyanobacterial host. (iii) During the past year, we have made great progress on a collaborative project with G. Marius Clore which involves designing, constructing and characterizing gp41-inspired fusion inhbitors. In addition to constructing Nccg-gp41, a nanomolar fusion inhibitor, we designed and constructed several stable trimeric inhibitors that incorporate the N-helices of native gp41. In addition to succeeding in inhibitor design, we constructured a mutant version refered to as N36-mut(e,g), so called because the 'e' and 'g' positions of the helical wheel, which are the amino acids necessary for formation of the putative fusogenic 6-helix bundle, have been mutated to abolish such interactions yet maintain interactions leading to intermolecular N-peptide trimer formation. Because this peptide potently inhibits fusion, these results indicate that a trimer of N-peptides of gp41 must exist in the fusion-competent form of HIV-1 Env. Antibodies have been generated to most of these constructs and are being evaluated as fusion inhibitors.